System and Method for Applying Machine Learning Algorithms to Compute Health Scores for Workload Scheduling

1. A method comprising: collecting, via at least one processor, first temporal statistics for a port element in a computing environment; collecting, via the at least one processor, second temporal statistics for a switch element in the computing environment; collecting, via the at least one processor, third temporal statistics for the computing environment, the collecting of the third temporal statistics including collecting historical and current metrics for the computing environment; computing, via the at least one processor, a spatial correlation between network features and network elements comprising the port element and the switch element; computing a health score for the port element or the switch element by factoring the first temporal statistics, the second temporal statistics, the third temporal statistics, and the spatial correlation; computing, via a machine learning technique, a port dynamic weight for the port element and a switch dynamic weight for the switch element; and scheduling a workload to consume computing resources in the computing environment based on the health score, the port dynamic weight, and the switch dynamic weight.

2. The method of claim 1 , wherein the collecting of the first temporal statistics includes collecting historical and current metrics for the port element.

3. The method of claim 1 , wherein the collecting of the second temporal statistics includes collecting historical and current metrics for the switch element.

4. The method of claim 1 , wherein the computing of the health score for one of the port element and the switch element includes using an equation to compute the health score as follows: H f t = ( ∑ i = 1 n ⁢ ( W i * P i ) * ( ∏ j = 1 m ⁢ R j t ) ) * ( ∏ k = 1 n ⁢ W k * P k ) , wherein: H is a health of a respective network element “f” at a time “t”; W is a weight calibrated using a neural network model for each network element; P is a normalized numeric value for each respective fabric element at the time “t”; (Π j=1 m R j t ): represents spatial dependencies of other attributes on a respective network element; “R” is a reward computed using reinforcement learning, for the spatial dependencies on the respective network element and multiplied to accommodate the dependency of other attributes; (Π k=1 n W k *P k ): where “P” is a normalized boolean value for each attribute and multiplied into the equation.

5. A system comprising: at least one processor; and a computer-readable storage device storing instructions which, when executed by the at least one processor, cause the at least one processor to perform operations comprising: collecting first temporal statistics for a port element in a computing environment; collecting second temporal statistics for a switch element in the computing environment; collecting third temporal statistics for the computing environment, the collecting of the third temporal statistics including collecting historical and current metrics for the computing environment; computing a spatial correlation between network features and network elements comprising the port element and the switch element; computing a health score for the port element or the switch element by factoring the first temporal statistics, the second temporal statistics, the third temporal statistics, and the spatial correlation; computing, via a machine learning technique, a port dynamic weight for the port element and a switch dynamic weight for the switch element; and scheduling a workload to consume computing resources in the computing environment based on the health score, the port dynamic weight, and the switch dynamic weight.

6. The system of claim 5 , wherein the collecting of the first temporal statistics includes collecting historical and current metrics for the port element.

7. The system of claim 5 , wherein the collecting of the second temporal statistics includes collecting historical and current metrics for the switch element.

8. The system of claim 5 , wherein the computing of the health score for one of the port element and the switch element includes using an equation to compute the health score as follows: H f t = ( ∑ i = 1 n ⁢ ( W i * P i ) * ( ∏ j = 1 m ⁢ R j t ) ) * ( ∏ k = 1 n ⁢ W k * P k ) , wherein: H is a health of a respective network element “f” at a time “t”; W is a weight calibrated using a neural network model for each network element; P is a normalized numeric value for each respective fabric element at the time “t”; (Π j=1 m R j t ): represents spatial dependencies of other attributes on the respective network element; “R” is a reward computed using reinforcement learning, for all the spatial dependencies on the respective network element and multiplied to accommodate the dependency of other attributes; (Π k=1 n W k *P k ): where “P” is a normalized boolean value for each attribute and multiplied into the equation.

9. A non-transitory computer-readable storage device storing instructions which, when executed by at least one processor, cause the at least one processor to perform operations comprising: collecting first temporal statistics for a port element in a computing environment; collecting second temporal statistics for a switch element in the computing environment; collecting third temporal statistics for the computing environment, the collecting of the third temporal statistics including collecting historical and current metrics for the computing environment; computing a spatial correlation between network features and network elements comprising the port element and the switch element; computing a health score for the port element or the switch element by factoring the first temporal statistics, the second temporal statistics, the third temporal statistics, and the spatial correlation; computing, via a machine learning technique, a port dynamic weight for the port element and a switch dynamic weight for the switch element; and scheduling a workload to consume computing resources in the computing environment based on the health score, the port dynamic weight, and the switch dynamic weight.

10. The non-transitory computer-readable storage device of claim 9 , wherein the collecting of the first temporal statistics includes collecting historical and current metrics for the port element.

11. The non-transitory computer-readable storage device of claim 9 , wherein the collecting of the second temporal statistics includes collecting historical and current metrics for the switch element.

12. The non-transitory computer-readable storage device of claim 9 , wherein the computing of the health score for one of the port element and the switch element includes using an equation to compute the health score as follows: H f t = ( ∑ i = 1 n ⁢ ( W i * P i ) * ( ∏ j = 1 m ⁢ R j t ) ) * ( ∏ k = 1 n ⁢ W k * P k ) , wherein: H is a health of a respective network element “f” at a time “t”; W is a weight calibrated using a neural network model for each network element; P is a normalized numeric value for each respective fabric element at the time “t”; (Π j=1 m R j t ): represents spatial dependencies of other attributes on the respective network element; “R” is a reward computed using reinforcement learning, for all the spatial dependencies on the respective network element and multiplied to accommodate the dependency of other attributes; (Π k=1 n W k *P k ): where “P” is a normalized boolean value for each attribute and multiplied into the equation.